LLMpediaThe first transparent, open encyclopedia generated by LLMs

VAXcluster

Generated by GPT-5-mini
Note: This article was automatically generated by a large language model (LLM) from purely parametric knowledge (no retrieval). It may contain inaccuracies or hallucinations. This encyclopedia is part of a research project currently under review.
Article Genealogy
Parent: DEC Hop 3
Expansion Funnel Raw 54 → Dedup 8 → NER 6 → Enqueued 0
1. Extracted54
2. After dedup8 (None)
3. After NER6 (None)
Rejected: 2 (not NE: 2)
4. Enqueued0 (None)
VAXcluster
NameVAXcluster
DeveloperDigital Equipment Corporation
Released1984
Discontinued1994 (product line transitions)
Latest releaseV5.0 (approx.)
Written inVAX/VMS, C, MACRO-32
Operating systemOpenVMS
PlatformVAX
LicenseProprietary
WebsiteNone

VAXcluster VAXcluster was a clustering technology introduced by Digital Equipment Corporation that provided high-availability and distributed resource management for VAX systems running OpenVMS. It enabled multiple Digital Equipment Corporation servers to present shared storage, coordinated process and file access, and failover capabilities to users and applications across sites. The system influenced later clustering products from Compaq, HP, and informed designs in projects at Sun Microsystems, IBM, and within research at Carnegie Mellon University and MIT.

Overview

VAXcluster combined hardware and software to create a single-system image across multiple Digital Equipment Corporation VAX computers, integrating components of OpenVMS, network protocols such as DECnet, and storage technologies like RAID arrays and SCSI. It supported features including distributed lock management, quorum rules, and rolling upgrades to minimize downtime—features later echoed in clustering solutions from Compaq and Hewlett-Packard after acquisitions. Key users included organizations such as NASA, Department of Defense (United States), AT&T, and universities like Stanford University and University of Cambridge leveraging clusters for critical services.

Architecture

The architecture centered on tightly-coupled VAX nodes interconnected via networks such as DECnet Phase IV, Ethernet standards including IEEE 802.3, and specialized interconnects like VAXcluster Interconnect (CI). Storage was shared through technologies compatible with Digital Storage Systems (DSS), SCSI controllers, and network-attached devices supported by Digital Equipment Corporation peripherals. Core architectural elements included the distributed lock manager (for coordination), global resource databases (for namespace consistency), and quorum algorithms to avoid split-brain conditions—concepts that parallel mechanisms in later systems from Sun Microsystems (Solaris Cluster), IBM (AIX Cluster), and Oracle Real Application Clusters. High-availability behaviors were achieved by integrating OpenVMS process migration tools, volume shadowing, and cluster-aware device drivers developed by teams at Digital Equipment Corporation.

Software and Implementation

Implementation relied on modifications and extensions to OpenVMS system services, cluster-aware file systems, and device drivers authored in languages such as C and MACRO-32. The software stack implemented global sections, distributed locking, and clusterwide event notification compatible with management consoles from DEC and third-party vendors like Tivoli and Legato Systems. VAXcluster interoperated with network services including DECnet, TCP/IP stacks of the era, and directory services used in deployments at Los Alamos National Laboratory and Sandia National Laboratories. Commercial software ported to clusters included databases and transaction systems used by Bank of America, scientific packages employed at European Organization for Nuclear Research (CERN), and mail/database services at Bell Labs.

Administration and Management

Cluster administration used OpenVMS management tools and DEC-supplied utilities for node addition/removal, backup, and configuration, often managed by system administrators with expertise from training programs at Digital Equipment Corporation and professional services from companies like Andersen Consulting (now Accenture). Management tasks included quorum configuration, device mapping, and shadow set maintenance; these were coordinated via cluster management consoles and command-line utilities common in environments at MITRE Corporation and RAND Corporation. Integration with third-party monitoring suites from HP and BMC Software assisted in operational oversight in enterprise contexts such as Federal Reserve System data centers.

Use Cases and Deployments

VAXcluster was deployed for high-availability computing, online transaction processing, and scientific computing at institutions including NASA Ames Research Center, Argonne National Laboratory, and commercial data centers run by MCI Communications Corporation. Typical use cases included database clustering, file serving with consistent locks across nodes for organizations like CERN and telecommunications switching support for carriers such as British Telecom and NTT. The technology was also used in fault-tolerant services for government installations such as National Security Agency research labs and academic computing clusters at University of California, Berkeley.

Compatibility and Legacy Issues

Compatibility relied on OpenVMS versions and specific VAX hardware families; transitions to Alpha and later Itanium hardware under Compaq and Hewlett-Packard required reengineering of cluster services, leading to product lines such as OpenVMS AlphaClusters. Legacy challenges included dependency on proprietary interconnects, driver availability, and migration of clustered storage to modern SAN and NAS technologies from vendors like EMC Corporation and NetApp. Preservation efforts by communities and institutions including Computer History Museum, university archives, and enthusiast groups have focused on documentation and emulator projects inspired by work at Free Software Foundation and research labs, while lessons from VAXcluster influenced clustering features in modern systems from Microsoft (Windows Server Failover Clustering), Red Hat (Pacemaker/Corosync integrations), and Oracle.

Category:Computer clusters